September_AMP_Digital

A D V A N C E D M A T E R I A L S & P R O C E S S E S | S E P T E M B E R 2 0 2 0 1 9 four “wave” defects are clearly shown in the same orientation as the embed- dedmarcelling and pillow insert defects (Fig. 5). The dry layer defects were not as clearly shown in the image at a selec- ted frequency range shown in Figure 4. The two red rectangular indications on each side of the scan image are alumi- num rails under the sample and are not embedded defects. The ability to see these features helps validate that a full volumetric inspection of the part was performed despite the dense com- posite structure. The collection of data in a multi-frequency mode allows one to select different frequency ranges to display the data. At a different selected frequency range (shown in Fig. 5), the “C” scan inspection image shows addi- tional dry defects in the part versus the first frequency range in Figure 4. The four wave features remain very clear in this frequency range. This demon- strates the superior analysis capability of a multi-frequency approach versus a single frequency inspection approach. Data collection over a range of fre- quencies allows the application of data analytics to convert the frequency- based signals to time-based signals. Thus, each of the data display modes of real, imaginary, magnitude, and phase can also be displayed versus time. The time-based display represents the time for the signal to travel from the trans- mitter, reflect from the indication, and return to the receiver. Thus, depth de- tection capability exists for a data set collected over a range of frequencies versus a single frequency. This method is similar to that used in some UT tech- niques but is a new area of application for MW technology. This depth detection methodolo- gy, while clearly viable, is not yet com- pletely refined and vetted to provide accurate depth measurement. Over- all accuracy is complicated as a result of the possibility of variable dielectric constant in complex materials, the air gap between the material surface and antenna, and the relative difference in magnitude of the speed of light and typical article thickness. Ongoing soft- ware improvements and data analysis techniques are expected to refine this approach with additional research en- abling a powerful tool for inspection. Not all defects are easily resolved using MW. For example, one challenge occurs in fiberglass when the sample manufacturer creates voids or delami- nations in the part. Voids or delamina- tions have a dielectric constant close to 1.0 (vacuum or air) while fiberglass has a dielectric constant of approximately 4.4. This makes those types of defects simple to detect using microwave in- spection. However, to simulate these defects, common methods use differ- ent styles of tape or even more sophisti- cated means, such as micro-balloons. If the tape or micro-balloon material has a dielectric close to that of fiberglass or is placed in a material that has a di- electric close to the defects, it will make the sample difficult to image with mi- crowave, although it may be easily de- tectable with UT or other methods. In this research, a sample contained mi- cro-balloon, tape, and pillow insert defects in a layer of adhesive. MW was unable to image the defects in this sam- ple. This is likely because these types of defects have a dielectric that is similar to the adhesive matrix that the defects were contained in. Therefore, proper development of representative sam- ples with particularly close attention Fig. 5 — Demonstration of MW inspection for defect sizing (depth detection) by using multi-frequency sweeping. Fig. 6 — Vector network analyzer with specially designed antennae capable of utilizing 6-14 GHz, fromCopper Mountain Technologies.